Aug 2026· International Conference on Information Security and Cryptology· pp. 1425-1431· 0 citations· 12 references
Abstract
The imbalance in loan records, the absence of data sharing opportunities, and the increased privacy regulations are becoming more problematic in terms of helping the financial institutions to assess credit risk. The paper presents a federated learning model that allows different institutions to create a common prediction model without access to the raw customer data. The model combines generative oversampling based on WGAN-GP with a Transformer classifier, the combination of which is the most significant novelty of the research as it contributes to the enhanced learning of minority defaults and ensures high locality of data. Experiments with the Give Me Some Credit dataset of 150,000 samples indicate that the proposed algorithm reaches a global AUC of 0.8405 and the method has a higher recall to identify high-risk borrowers than the traditional centralized training. These findings indicate that the framework supports prediction reliability as well as favors both law and privacy concerns in financial settings.
The results indicate that the proposed edge-driven federated learning framework can support privacy-preserving and robust cross-institutional financial risk modeling and provides an effective solution for collaborative fraud detection and anti-money laundering under data isolation, heterogeneous edge environments, and...
Wan-Li Zhang· ICST Transactions on Scalabl...· 0 citations
This paper investigates a privacy-preserving approach to churn prediction that combines federated learning (FL) with differential privacy (DP), and highlights the potential of privacy-preserving federated learning for practical distributed analytics applications where protecting sensitive data is essential.
The growing number of problems related to the privacy and security of data in cloud computing requires the need for efficient protection schemes. In this paper, we present an innovative solution for ensuring the privacy and management of data in large language models deployed in clouds by applying risk mitigation techn...
This dissertation proposes and evaluates DP-FedSHAP, a new architecture that applies client-level differential privacy only to post-hoc TreeSHAP vectors and measures the trade-off between explanation fidelity, privacy preservation, and the model's Area Under the Precision-Recall Curve (AUPRC).
FedHDL (Federated Heterogeneous Deep Learning), a novel privacy-preserving framework for cryptocurrency fraud detection that enables collaborative model training across five heterogeneous institutional nodes without raw data exchange, is introduced.
Kanika Singhal· Journal of Intelligent Decis...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
